I guess the Semi-Slav is statistically accurate enough. Any lines of QGD or QGA you would like me to investigate?
Let's proceed step by step. 
First of all, could U please post statistical evaluations of black replies to 1. d4 as I've done for 1. e4 in the post #4 ? Statistically correct replies are supposed to have evaluations =< 0.5.
I'm a bit tired at the moment, going to watch Kill la Kill anime. 
I'll come back later.
Differentiation2: Great, thanks! 
I'll think about results a bit later. If U are motivated, please post evaluation results for any QG lines of your choice.
P.S. I feel that we'll be obliged to create other threads to discuss statistic evaluations of main lines of most popular openings.
Yigor: It's completely fine if we have to make more threads at least for me. I have more than enough time, ex: I am also doing a study with a friend about chess openings that is suppost to be ~64 chapters long, each chapter being one opening. That means the Sicilian, Spanish, French, even the Grob (ok, not the Grob. Grob for most theory!!) all have their own chapters instead of Dragon, Najdorf, Sicilian Variation 1000000042, etc.
Great! I have also enough free time during this summer.
1 1...d6: +0.218 (Strange, because after e4 we are at Pirc)
Interesing discovery! It seems to be just the best statistical reply to 1. d4. 
I created a special thread
https://www.chess.com/forum/view/chess-openings/lowe-s-defense
calling it Löwe's defense.
I created a new thread about the subcategory of statistically excellent openings :
https://www.chess.com/forum/view/chess-openings/statistically-excellent-openings
Optimissed: We don't believe that the statistics is omnipotent and can answer all questions. However, if Slav is really so solid, why white wins almost 2 times more often than black (in 63,000+ master games) ?!?
So it seems that 1...g6 is Statistically Excellent against 1.e4, 1.d4, 1.Nf3, and 1.c4, all the main openings.
Interesting and unexpected result.
Differentiation2: Could U please calculate and post (like #4 or #37) today evaluation results for 1. Nf3 (or 1. c4) ? and I'll post results for 1. c4 (or 1. Nf3). We shouldn't forget 1...g3 too.
I will start on 1.c4
Great! I'll just watch an episod of Claymore anime and will post results of English opening in about 1 hour.
Very interesting indeed. I have my own explanation of why d6 and g6 are better than other defenses for black. Can you calculate for the first two moves like e4 and d4 for white and g6 and d6 for black. Also d4 and c4 replied with g6 and d6. What about e4 and d4 replied with b6 and e6 or even d4 and c4 with b6 and e6.
I will share my thoughts about these openings later.
Including both d6 and g6 or including both b6 and e6 for the first two moves should be better for black than any other first two moves according to my thinking. Let us see the statistics first to confirm if it is right or not.
My understanding is that the openings which are highly tactical and double edged are generally more drawish because after the end of the tactical phase, the positions tend to become dull and drawish in the late middlegame or endgame. But on the other hand, the more positional openings at the start will explode with tactics later on in the middlegame and this makes blunders in middlegames highly likely leading into decisive results for either white or black. What do you say about this? It is my observation and I don't have any statistical proof for this.
This explains why Traxler is more drawish than BDG!
One more thing that I have noticed is symmetry is generally bad for black. Black does best when having counter chances. That is one reason why Sicilian, Modern, Indian defenses are better for black than QGD and Double king pawn openings.
My another observation is that hyper-modern defenses are better for black. Thats why I think Modern, Pirc, QID, KID look statistically better than Sicilian, QGD, Slav, French, Caro Cann.
The possible explanation is that black avoiding early exchanges of pawns and then striking later at the white central pawns is more dangerous for white because maintaining the center becomes very difficult as the game goes on due to poor positioning of pieces and a possible over-extension of white pawns in the center. One more reason is that it gives more options for white to choose a wrong plan when he has got lot of options to continue.
One more big surprise: 1...d5 is statistically incorrect reply to the Queen's Pawn opening 1. d4, ev(1...d5)=+0.7.
Wow! This was almost as unexpected as Rubinstein. I will look into the Slav, Semi-Slav, QGA and QGD to see if it is just the inaccurate moves messing this up.
Good idea.